大数据云环境下TDS和BUG混合k-匿名化方法

doi:10.11959/j.issn.1000-0801.2016135

摘要/Abstract

摘要：

针对一般子树匿名化方法处理大数据效率低和伸缩性较差的问题，提出了一种可伸缩的自下向上的泛化（BUG）方法，并在此基础上，结合已有的自上向下的特化（TDS），形成一种混合方法。在提出的方法中，k-匿名作为隐私模型，TDS和BUG都是基于映射化简开发组成，并通过云的强大计算能力来获得较高的伸缩性。提出的映射化简BUG只需在几次泛化循环之后就可插入一个新的泛化候选，不会影响另一个泛化的信息损失。考虑到工作负载平衡点K与匿名参数k的复杂关系，将映射化简的BUG和TDS结合形成混合方法。实验结果验证了本文方法的有效性，与TDS和BUG相比，混合方法的效率和可伸缩性大为提高。

关键词: 云计算, 子树匿名化, 大数据, 泛化, 特化, 映射化简

Abstract:

As the issue of low efficiency and poor scalability in general sub-tree anonymous method of treating big data,a bottom-up generalization(BUG) method with scalability was proposed,and on this basis,combined with the existing top-down specialization(TDS),a hybrid approach was formed.In the proposed method,k-anonymity was being as a privacy model,the compositions of TDS and BUG were developed with mapping simplification,and higher scalability through powerful cloud computing capabilities were achieved.The proposed mapping simplification BUG could insert a new candidate after several cycles of generalization,and would not affect information loss of another generalization.Given the complexity of the relationship between workload balancing point K and anonymous parameter k,mapping simplifications of BUG and TDS were combined to form a hybrid approach.Experimental results demonstrate the effectiveness of the proposed method and compared with TDS and BUG,the efficiency and scalability of hybrid method are greatly improved.

Key words: cloud computing, sub-tree anonymous, big data, generalization, specialization, mapping simplification

范晓峰,闫凤,刘洋. 大数据云环境下TDS和BUG混合k-匿名化方法[J]. 电信科学, 2016, 32(7): 90-96.

Xiaofeng FAN,Feng YAN,Yang LIU. Hybrid k-anonymity approach based on TDS and BUG under the environment of big data cloud[J]. Telecommunications Science, 2016, 32(7): 90-96.

图/表 3

参考文献 17

[1]	戚建国 . 基于云计算的大数据安全隐私保护的研究[D]．北京：北京邮电大学， 2015. QI J G . Research on security privacy protection of large data based on cloud computing[D]. Beijing: Beijing University of Posts and Telecommunications, 2015.
[2]	康瑛石，郑子军，大数据整合机制与信息共享服务实现[J]. 电信科学， 2014,30(12): 97-102. KANG Y S , ZHENG Z J . Big data integration mechanism and information sharing service realization[J]. Telecommunications Science, 2014,30(12): 97-102.
[3]	徐勇，秦小麟，杨一涛，等. 一种考虑属性权重的隐私保护数据发布方法[J]. 计算机研究与发展， 2012,49(5): 913-924. XU Y , QIN X L , YANG Y T , et al. A QI weighted-aware approach to privacy preserving publishing data set[J]. Journal of Computer Research and Development, 2012,49(5): 913-924.
[4]	MACHANAVAJJHALA A , GEHRKE J , KIFER D , et al. L-diversity: privacy beyond k-anonymity[J]. ACM Transactions on Knowledge Discovery from Data, 2007,1(1): 24-32.
[5]	李明庆 . 一种基于抽样确定泛化区间的K-匿名算法[J]. 哈尔滨：哈尔滨工程大学， 2013. LI M Q . A K-anonymity algorithm based on sampling to determine the generalized interval[D]. Harbin: Harbin Engineering University, 2013.
[6]	WANG K , FUNG B C M , YU P S . Handicapping attacker's confidence: an alternative to k-anonymization[J]. Knowledge ＆Information Systems, 2007,11(3): 345-368.
[7]	TERROVITIS M , MAMOULIS N , LIAGOURIS J , et al. Privacy preservation by disassociation[J]. Proceedings of the Vldb Endowment, 2012,5(10): 944-955.
[8]	黄春梅，费耀平，李敏 , 等. 基于多维泛化路径的K-匿名算法[J]. 计算机工程， 2009,35(2): 154-156. HUANG C M , FEI Y P , LI M . K-anonymity algorithm based on multi-dimensional generalization path[J]. Computer Engineering, 2009,35(2): 154-156.
[9]	FUNG B C M , WANG K , YU P S . Anonymizing classification data for privacy preservation[J]. IEEE Transactions on Knowledge ＆ Data Engineering, 2007,19(5): 711-725.
[10]	RAJALAKSHMI V , MALA G S A . Anonymization based on nested clustering for privacy preservation in data mining[J]. Indian Journal of Computer Science ＆ Engineering, 2013,34(3): 856-861.
[11]	MOHAMMED N , FUNG B C M , HUNG P C K , et al. Centralized and distributed anonymization for high-dimensional healthcare data[J]. ACM Transactions on Knowledge Discovery from Data, 2010,4(4): 885-900.
[12]	ZHANG X , YANG L T , LIU C , et al. A scalable two-phase top-down specialization approach for data anonymization using MapReduce on cloud[J]. IEEE Transactions on Parallel ＆Distributed Systems, 2014,25(2): 363-373.
[13]	吕品，钟珞，于文兵 , 等. MA-Datafly：一种支持多属性泛化的k-匿名方法[J]. 计算机工程与应用， 2013,49(4): 138-140. LV P , ZHONG L , YU W B , et al. MA-Datafly: k anonymity approaches for supporting multi-attribute generalization[J]. Computer Engineering and Applications, 2013,49(4): 138-140.
[14]	刘盼盼 . 基于聚类的数据匿名发布技术研究[D]．西安：西安电子科技大学， 2013. LIU P P . Research on technology of data anonymous publishing based on clustering[D]. Xi'an: Xidian University, 2013.
[15]	CUI X , ZHU P , YANG X , et al. Optimized big data K-means clustering using map reduce[J]. Journal of Super Computing, 2014,70(3): 1249-1259.
[16]	王笑帝，张云勇，刘镝，等. 云计算虚拟化安全技术研究[J]. 电信科学， 2015,31(6): 1-5. WANG X D , ZHANG Y Y , LIU D , et al. Research on security of virtualization on cloud computing[J]. Telecommunications Science, 2015,31(6): 1-5.
[17]	KVM[EB/OL]. (2011-04-10)[2012-02-10]. .

符号	定义
D	包含数据记录的数据集
M	属性数量
r	数据记录，r∈D≤且r=<v₁,v₂,…,v_m,sv>，其中，v_i(1≤i≤m)为属性值，sv为敏感值
Attr_i	数据记录的第i个属性
TT_i	属性Attr_i的分类树
DOM_i	TT_i中所有域值的集合
SV	敏感值集合
qid	准标识符，qid=<q₁,q₂,…,q_m>,q_i∈DOM_i,1≤i≤m
QID	准标识符集QID=<Attr₁,Attr₂,…,Attr_m>
QID(qid)	具有相同准标识符qid的包含所有记录的准标识符组
AGSet	可用的泛化集
SGSet	同科泛化集
NGSet	新的泛化集
AQISet	匿名准标识符
CGSet	关键泛化集
NCGSet	非关键泛化集
RGSet	竞赛泛化集
ACGSet	可用关键泛化集